Re: Multiple servers in a Ignite Cluster
Tracy, You can limit the set of nodes where the cache is deployed via CacheConfiguration.setNodeFilter() config property. Generally, all your nodes should be in the same cluster, but you can create multiple roles and logical cluster groups. -Val -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Multiple-servers-in-a-Ignite-Cluster-tp8840p8974.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
Re: Multiple servers in a Ignite Cluster
Hi Vlad, The ideal work flow for my use case is: I host two clusters, one is computation cluster that run Spark jobs, the other is data cluster that host Ignite node and cache hot data. Then at the run time, multiple Spark jobs share this data cluster and query it. The problem I have is, I am pre-loading Ignite cache using a Spark job. Once IgniteContext got instantiated, it will launch Ignite Node with same number of Spark executors I allocated. Then distributed cache will distribute data on those nodes within my computation cluster as well, which I don't want because partial data hosted by these nodes won't be there once my pre-load job dies. Currently I force these node to be client mode so that the cache only distributed to data cluster when I execute my pre-load job. Are there better way to solve this? Thanks, Tracy -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Multiple-servers-in-a-Ignite-Cluster-tp8840p8887.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
Re: Multiple servers in a Ignite Cluster
Thanks. Works fine now. -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Multiple-servers-in-a-Ignite-Cluster-tp8840p8851.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
Re: Multiple servers in a Ignite Cluster
HI, It will not always. Discovery SPI responsible for this process. It will be may TcpDiscoveryMulticastIpFinder[1] (any node with same configuration and starts in same network will be joined) or TcpDiscoveryVmIpFinder[2] (only nodes with specific IP will be joined) or any other. [1]: https://apacheignite.readme.io/v1.7/docs/cluster-config#multicast-based-discovery [2]: https://apacheignite.readme.io/v1.7/docs/cluster-config#isolated-ignite-clusters-on-the-same-set-of-machin On Wed, Nov 9, 2016 at 7:33 PM, Tracyl <tlian...@bloomberg.net> wrote: > Thanks. In that case, my question is how to define the scope of cluster(Or > how to specify the cluster a server belongs to)? I assume if someone else > start a ignite node, would my ignite server auto-discover it as well? > > > > -- > View this message in context: http://apache-ignite-users. > 70518.x6.nabble.com/Multiple-servers-in-a-Ignite-Cluster-tp8840p8844.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com. > -- Vladislav Pyatkov
Re: Multiple servers in a Ignite Cluster
Thanks. In that case, my question is how to define the scope of cluster(Or how to specify the cluster a server belongs to)? I assume if someone else start a ignite node, would my ignite server auto-discover it as well? -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Multiple-servers-in-a-Ignite-Cluster-tp8840p8844.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
Re: Multiple servers in a Ignite Cluster
Hi Tracyl, You are need to start each Ignite service separately. But you always can do it over ssh using simple bash script. On Wed, Nov 9, 2016 at 6:55 PM, Tracyl <tlian...@bloomberg.net> wrote: > I have following ignite config: > > def initializeIgniteConfig() = { > val ipFinder = new TcpDiscoveryVmIpFinder() > > val HOST = "xx.xx.xx.xx:47500..47509" > ipFinder.setAddresses(Collections.singletonList(HOST)) > > val discoverySpi = new TcpDiscoverySpi() > discoverySpi.setIpFinder(ipFinder) > > val igniteConfig = new IgniteConfiguration() > igniteConfig.setDiscoverySpi(discoverySpi) > > //Ignite uses work directory as a relative directory for internal > write activities, for example, logging. > // Every Ignite node (server or client) has it's own work directory > independently of other nodes. > > igniteConfig.setWorkDirectory("/tmp") > igniteConfig > } > > My use case is: I would like to have multiple ignite servers, each server > cache a subset of the data and I send distributed closures to each node to > do local computation. In this case, can I start multiple servers on one > single machine by just passing multiple ip and port? Or I will need to > start > each server on each machine separately? Thanks in advance! > > > > -- > View this message in context: http://apache-ignite-users. > 70518.x6.nabble.com/Multiple-servers-in-a-Ignite-Cluster-tp8840.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com. > -- Vladislav Pyatkov
Multiple servers in a Ignite Cluster
I have following ignite config: def initializeIgniteConfig() = { val ipFinder = new TcpDiscoveryVmIpFinder() val HOST = "xx.xx.xx.xx:47500..47509" ipFinder.setAddresses(Collections.singletonList(HOST)) val discoverySpi = new TcpDiscoverySpi() discoverySpi.setIpFinder(ipFinder) val igniteConfig = new IgniteConfiguration() igniteConfig.setDiscoverySpi(discoverySpi) //Ignite uses work directory as a relative directory for internal write activities, for example, logging. // Every Ignite node (server or client) has it's own work directory independently of other nodes. igniteConfig.setWorkDirectory("/tmp") igniteConfig } My use case is: I would like to have multiple ignite servers, each server cache a subset of the data and I send distributed closures to each node to do local computation. In this case, can I start multiple servers on one single machine by just passing multiple ip and port? Or I will need to start each server on each machine separately? Thanks in advance! -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Multiple-servers-in-a-Ignite-Cluster-tp8840.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.